pith. sign in

arxiv: 2501.13188 · v2 · pith:SC5STYSWnew · submitted 2025-01-22 · ❄️ cond-mat.stat-mech · cs.LG· nlin.AO· q-bio.CB

Topological constraints on self-organisation in locally interacting systems

Francesco Sacco , Dalton A R Sakthivadivel , Michael Levin This is my paper

Pith reviewed 2026-05-23 05:27 UTC · model grok-4.3

classification ❄️ cond-mat.stat-mech cs.LGnlin.AOq-bio.CB
keywords topological constraintsself-organisationdomain wallsPotts modelhierarchical networksspontaneous orderinggraph combinatoricscollective intelligence
0
0 comments X

The pith

The combinatorics of interactions on a planar graph impose necessary conditions for spontaneous ordering to occur.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that systems of locally interacting parts on the vertices of a planar graph can exhibit long-range order only when the graph's topology satisfies certain conditions. It examines this by tracking how free energy scales when domain walls form in the Potts model, autoregressive models, and hierarchical networks. A sympathetic reader cares because the result constrains when aligned parts can sustain a system-level ordered state, directly bearing on collective intelligence in living systems and engineered models. The central mechanism is the way edge combinatorics raise or lower the cost of breaking global alignment into separate domains.

Core claim

In a system whose interactions follow the edges of a planar graph, the free-energy penalty incurred by inserting a domain wall scales differently according to the graph's combinatorial structure; topologies that make this penalty grow sufficiently fast with system size permit an ordered phase, while others forbid it. This holds across the Potts model on regular lattices, autoregressive sequence models, and hierarchical networks, and it explains why multiscale biological architectures support complex patterning whereas flat language-model interaction graphs do not.

What carries the argument

Scaling of free energy cost under domain-wall insertion, which encodes the topological constraints that allow or prevent an ordered phase.

If this is right

  • Biological multiscale architectures can reach complex ordered states because their hierarchical interaction graphs make domain-wall formation costly.
  • Rudimentary language models fail on long coherent outputs because their interaction graphs keep domain walls cheap.
  • Any locally interacting system must satisfy the same topological conditions on its interaction graph to maintain an ordered target state.
  • Spontaneous ordering is impossible on graphs whose combinatorics allow domain walls to proliferate without free-energy penalty.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Engineering the topology of an interaction graph could be used to enforce or suppress global order in artificial collectives.
  • The same domain-wall argument may apply to other collective systems such as neural populations or social networks whose wiring diagrams are known.
  • If the scaling relation holds, then rewiring a flat graph into a hierarchical one should restore long-range order in an otherwise disordered model.

Load-bearing premise

The free-energy scaling observed under domain-wall formation in the three chosen models directly encodes the topological constraints that matter for self-organization in biology and language models.

What would settle it

A simulation or analytic calculation on a hierarchical graph that the paper predicts should support ordering, yet in which domain walls remain energetically cheap at large size and no ordered phase appears.

Figures

Figures reproduced from arXiv: 2501.13188 by Dalton A R Sakthivadivel, Francesco Sacco, Michael Levin.

Figure 1
Figure 1. Figure 1: A local Hamiltonian in dimension one. The Hamiltonian of this chain is a sum of windowed Hamiltonians of length ω = 5. Pictured is the second window; the first begins at s−1. Our methodology will be given by the following prescription: 1. Begin with the system in one of the ordered configurations, e.g. one of the stored patterns 2 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Two different domains in a two-dimensional grid. The perimeter that separates the two domains is drawn in orange. Let H be a graph Hamiltonian and P be the peri￾meter length of a domain wall. As the perimeter length increases, the number of possible configur￾ations of domain barriers increases, increasing the entropy of the system ∆S. We say that the entropy gained scales as fS if ∆S = O(fS(P)). We give a … view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of a graph Hamiltonian. Left: [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: Text generation in analogy to a morphogenetic process. We will now discuss a similar result as the one in the previous subsection, for autoregressive models. 5 [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: An autoregressive model with ω = 5. To fit this in the discussion of patterns and long￾range order in spin chains, we find the following theorem useful. We will set the convention that if u − ω < 1 then the window is empty at that index, and that conditioning on the empty set is the same as taking unconditional probability. Theorem 3. A unique local Hamiltonian with win￾dow length ω can be associated to an… view at source ↗
Figure 6
Figure 6. Figure 6: A graph G with two independent 3-vertex cliques and two edges connecting the cliques. For brevity, and without loss of generality, we will assume the coupling constants J are uniform across the graph. We will need the following observation before we prove our main theorem in this section. Recall that Boltzmann’s formula S = k log W de￾notes by W the multiplicity of a macrostate. A clique is positively (neg… view at source ↗
Figure 7
Figure 7. Figure 7: Seven independent cliques are highlighted [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Cognition as goal-directed behaviour, [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
read the original abstract

All intelligence is collective intelligence, in the sense that it is made of parts which must align with respect to system-level goals. Understanding the dynamics which facilitate or limit navigation of problem spaces by aligned parts thus impacts many fields ranging across life sciences and engineering. To that end, consider a system on the vertices of a planar graph, with pairwise interactions prescribed by the edges of the graph. Such systems can sometimes exhibit long-range order, distinguishing one phase of macroscopic behaviour from another. In networks of interacting systems we may view spontaneous ordering as a form of self-organisation, modelling neural and basal forms of cognition. Here, we discuss necessary conditions on the topology of the graph for an ordered phase to exist, with an eye towards finding constraints on the ability of a system with local interactions to maintain an ordered target state. By studying the scaling of free energy under the formation of domain walls in three model systems -- the Potts model, autoregressive models, and hierarchical networks -- we show how the combinatorics of interactions on a graph prevent or allow spontaneous ordering. As an application we are able to analyse why multiscale systems like those prevalent in biology are capable of organising into complex patterns, whereas rudimentary language models are challenged by long sequences of outputs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims that necessary conditions on the topology of planar graphs exist for ordered phases to arise in systems with local pairwise interactions on the vertices. By analyzing the scaling of free-energy costs under domain-wall formation in the Potts model, autoregressive models, and hierarchical networks, it argues that the combinatorics of interactions on the graph either prevent or allow spontaneous ordering. This framework is then applied to explain why multiscale biological systems can form complex patterns while rudimentary language models struggle with long output sequences.

Significance. If the central claim holds, the manuscript identifies topology-dependent constraints on self-organization that could inform models of collective intelligence across biology and machine learning. A strength is the comparative use of three distinct model classes (Potts, autoregressive, hierarchical) to probe the same question, which provides a broader basis than a single-system analysis. The work also attempts to connect statistical-mechanics scaling arguments to applied domains, though the transfer remains at a qualitative level.

major comments (2)
  1. [Abstract] Abstract: the central claim requires that free-energy scaling under domain-wall formation depends only on graph properties (planarity, cycle structure, degree sequence) and not on the specific Hamiltonian or factorization details of the three models. No indication is given that this separation has been performed; without it the argument yields model-specific behavior rather than general topological necessary conditions. This is load-bearing for the transfer to arbitrary locally interacting networks.
  2. [Abstract] Abstract (application paragraph): the assertion that the same topological constraints explain both biological multiscale organization and language-model difficulties with long sequences rests on the unshown premise that the three models capture the relevant graph combinatorics in those domains. No quantitative mapping or falsifiable prediction is supplied to support the transfer.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'combinatorics of interactions on a graph' is used without a preceding definition or example of the combinatorial object being counted; a short clarifying sentence would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the scope of our claims. We address each major point below and will revise the abstract to improve precision.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim requires that free-energy scaling under domain-wall formation depends only on graph properties (planarity, cycle structure, degree sequence) and not on the specific Hamiltonian or factorization details of the three models. No indication is given that this separation has been performed; without it the argument yields model-specific behavior rather than general topological necessary conditions. This is load-bearing for the transfer to arbitrary locally interacting networks.

    Authors: The three models were deliberately chosen to have distinct Hamiltonians and interaction factorizations (symmetric pairwise in Potts, directed sequential in autoregressive, and recursive tree-structured in hierarchical) while sharing the same underlying planar graphs. Explicit calculations in Sections 3–5 demonstrate that the leading free-energy scaling of domain walls is controlled by graph invariants such as cycle parity and planarity-induced embedding constraints, independent of the model-specific energy or entropy terms. We will revise the abstract to state explicitly that the necessary conditions are graph-theoretic and hold across these differing interaction structures. revision: yes

  2. Referee: [Abstract] Abstract (application paragraph): the assertion that the same topological constraints explain both biological multiscale organization and language-model difficulties with long sequences rests on the unshown premise that the three models capture the relevant graph combinatorics in those domains. No quantitative mapping or falsifiable prediction is supplied to support the transfer.

    Authors: The applications are presented as qualitative illustrations of how the identified topological constraints can manifest in different domains, using the three models as representative classes of locally interacting systems. We do not claim a quantitative mapping or direct falsifiable predictions at this stage. We will revise the abstract to qualify the biological and language-model examples as conceptual applications of the framework, noting that quantitative validation lies beyond the present scope. revision: partial

Circularity Check

0 steps flagged

Derivation self-contained with no circular steps

full rationale

The paper establishes necessary topological conditions for ordered phases by explicit computation of domain-wall free-energy scaling in three concrete models (Potts, autoregressive, hierarchical). These models are independently defined, and the extraction of graph combinatorics (planarity, cycles, etc.) from their scaling behavior constitutes an independent derivation rather than a reduction to fitted parameters, self-definitions, or prior self-citations. No load-bearing step equates the claimed topological constraint to its own inputs by construction; the argument remains falsifiable against the models' Hamiltonians and is not forced by renaming or ansatz smuggling.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, axioms, or invented entities can be extracted or verified from the provided text.

pith-pipeline@v0.9.0 · 5764 in / 992 out tokens · 26565 ms · 2026-05-23T05:27:32.596589+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

69 extracted references · 69 canonical work pages · 2 internal anchors

  1. [1]

    Topological constraints on self-organisation in locally interacting systems

    or diffusion models [24]. In these systems, in ∗ DARS is supported by the Einstein Chair programme at the CUNY Graduate Center and the VERSES Research Lab. † dsakthivadivel@gc.cuny.edu order to maintain a pattern out of equilibrium or produce sensible data over long time spans, there must exist a ‘condensed’ phase with long-range or- der. Whilst some syst...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  2. [2]

    one of the stored patterns 2

    Begin with the system in one of the ordered configurations, e.g. one of the stored patterns 2

    work page

  3. [3]

    Create a domain wall

    work page

  4. [4]

    Estimate the energy gained by the system

    work page

  5. [5]

    Instead we can use the structure imposed by the windows

    Find the asymptotics of the free energy as the number of domain walls increases Computing the change in F by changing the num- ber of domain walls requires detailed knowledge of the combinatorics of the interactions on the graph. Instead we can use the structure imposed by the windows. Namely, we have regulated the length of interactions such that we need...

    work page

  6. [6]

    Collective minds

    Iain D Couzin. Collective minds. Nature, 445(7129):715–715, 2007

    work page 2007

  7. [7]

    self-modeling

    Richard A Watson, Christopher L Buckley, and Rob Mills. Optimization in “self-modeling” complex ad- aptive systems. Complexity, 16(5):17–26, 2011

    work page 2011

  8. [8]

    The computational boundary of a “self”’: developmental bioelectricity drives multicel- lularity and scale-free cognition

    Michael Levin. The computational boundary of a “self”’: developmental bioelectricity drives multicel- lularity and scale-free cognition. Frontiers in Psy- chology, 10:2688, 2019

    work page 2019

  9. [9]

    A scalable pipeline for design- ing reconfigurable organisms

    Sam Kriegman, Douglas Blackiston, Michael Levin, and Josh Bongard. A scalable pipeline for design- ing reconfigurable organisms. Proceedings of the National Academy of Sciences , 117(4):1853–1859, 2020

    work page 2020

  10. [10]

    Design for an individual: connec- tionist approaches to the evolutionary transitions in individuality

    Richard A Watson, Michael Levin, and Chris- topher L Buckley. Design for an individual: connec- tionist approaches to the evolutionary transitions in individuality. Frontiers in Ecology and Evolution , 10:823588, 2022

    work page 2022

  11. [11]

    Cellular and evolutionary perspectives on organismal cognition: from unicellular to multicel- lular organisms

    Frantiˇ sek Baluˇ ska, William B Miller, and Arthur S Reber. Cellular and evolutionary perspectives on organismal cognition: from unicellular to multicel- lular organisms. Biological Journal of the Linnean Society, 139(4):503–513, 2023

    work page 2023

  12. [12]

    A revised central dogma for the 21st century: all biology is cognitive information processing

    William B Miller, Frantiˇ sek Baluˇ ska, and Arthur S Reber. A revised central dogma for the 21st century: all biology is cognitive information processing. Pro- gress in Biophysics and Molecular Biology , 2023

    work page 2023

  13. [13]

    The spirit of D’Arcy thompson dwells in empirical morphospace

    Jonathon R Stone. The spirit of D’Arcy thompson dwells in empirical morphospace. Mathematical Biosciences, 142(1):13–30, 1997

    work page 1997

  14. [14]

    Collect- ive behavior in cancer cell populations

    Thomas S Deisboeck and Iain D Couzin. Collect- ive behavior in cancer cell populations. Bioessays, 31(2):190–197, 2009

    work page 2009

  15. [15]

    The old and new faces of mor- phology: the legacy of D’Arcy thompson’s ‘theory of transformations’ and ‘laws of growth’

    Arhat Abzhanov. The old and new faces of mor- phology: the legacy of D’Arcy thompson’s ‘theory of transformations’ and ‘laws of growth’. Develop- ment, 144(23):4284–4297, 2017

    work page 2017

  16. [16]

    Cognition in some surprising places

    Arthur S Reber and Frantiˇ sek Baluˇ ska. Cognition in some surprising places. Biochemical and Biophysical Research Communications, 564:150–157, 2021

    work page 2021

  17. [17]

    Reframing cognition: getting down to biological basics

    Pamela Lyon, Fred Keijzer, Detlev Arendt, and Mi- chael Levin. Reframing cognition: getting down to biological basics. Philosophical Transactions of the Royal Society B, 376(1820):20190750, 2021

    work page 2021

  18. [18]

    Technological approach to mind everywhere: an experimentally-grounded frame- work for understanding diverse bodies and minds

    Michael Levin. Technological approach to mind everywhere: an experimentally-grounded frame- work for understanding diverse bodies and minds. Frontiers in Systems Neuroscience, 16:768201, 2022

    work page 2022

  19. [19]

    The scaling of goals from cellular to anatomical homeostasis: an evolutionary simulation, experiment and analysis

    L´ eo Pio-Lopez, Johanna Bischof, Jennifer V LaP- alme, and Michael Levin. The scaling of goals from cellular to anatomical homeostasis: an evolutionary simulation, experiment and analysis. Interface Fo- cus, 13(3):20220072, 2023

    work page 2023

  20. [20]

    Cellular signaling pathways as plastic, proto-cognitive systems: Implications for biomedicine

    Juanita Mathews, Alan Jaelyn Chang, Liam Devlin, and Michael Levin. Cellular signaling pathways as plastic, proto-cognitive systems: Implications for biomedicine. Patterns, 4(5), 2023

    work page 2023

  21. [21]

    Bioelectric networks: the cognitive glue enabling evolutionary scaling from physiology to mind

    Michael Levin. Bioelectric networks: the cognitive glue enabling evolutionary scaling from physiology to mind. Animal Cognition, pages 1–27, 2023

    work page 2023

  22. [22]

    On Purpose

    Michael Levin. Collective intelligence of morpho- genesis as a teleonomic process. In Evolution “On Purpose”: Teleonomy in Living Systems . The MIT Press, 08 2023

    work page 2023

  23. [23]

    Future medicine: from molecular pathways to the collective intelli- gence of the body

    Eric Lagasse and Michael Levin. Future medicine: from molecular pathways to the collective intelli- gence of the body. Trends in Molecular Medicine , 2023

    work page 2023

  24. [24]

    Neural networks and physical sys- tems with emergent collective computational abil- ities

    John J Hopfield. Neural networks and physical sys- tems with emergent collective computational abil- ities. Proceedings of the National Academy of Sci- ences, 79(8):2554–2558, 1982

    work page 1982

  25. [25]

    Dense associ- ative memory for pattern recognition

    Dmitry Krotov and John J Hopfield. Dense associ- ative memory for pattern recognition. Advances in Neural Information Processing Systems, 29, 2016

    work page 2016

  26. [26]

    On a model of associative memory with huge storage capacity

    Mete Demircigil, Judith Heusel, Matthias L¨ owe, Sven Upgang, and Franck Vermet. On a model of associative memory with huge storage capacity. Journal of Statistical Physics , 168:288–299, 2017

    work page 2017

  27. [27]

    Spin-glass theory for pedestrians

    Tommaso Castellani and Andrea Cavagna. Spin-glass theory for pedestrians. Journal of Statistical Mechanics: Theory and Experiment , 2005(05):P05012, 2005

    work page 2005

  28. [28]

    Efficient learning of sparse representations with an energy-based model

    Marc’Aurelio Ranzato, Christopher Poultney, Sumit Chopra, and Yann Le Cun. Efficient learning of sparse representations with an energy-based model. Advances in Neural Information Processing Systems, 19, 2006

    work page 2006

  29. [29]

    Diffusion models: A comprehensive survey of methods and applications

    Ling Yang, Zhilong Zhang, Yang Song, Shenda Hong, Runsheng Xu, Yue Zhao, Wentao Zhang, Bin Cui, and Ming-Hsuan Yang. Diffusion models: A comprehensive survey of methods and applications. ACM Computing Surveys , 56(4):1–39, 2023

    work page 2023

  30. [30]

    Simulation of biological cell sorting using a two-dimensional extended Potts model

    Fran¸ cois Graner and James A Glazier. Simulation of biological cell sorting using a two-dimensional extended Potts model. Physical Review Letters , 69(13):2013, 1992

    work page 2013

  31. [31]

    On multiscale approaches to three-dimensional model- ling of morphogenesis

    Rajiv Chaturvedi, Chengbang Huang, Bogdan Kazmierczak, T Schneider, Jesus A Izaguirre, 10 Tilmann Glimm, H George E Hentschel, J A Glazier, S A Newman, and M S Alber. On multiscale approaches to three-dimensional model- ling of morphogenesis. Journal of the Royal Society Interface, 2(3):237–253, 2005

    work page 2005

  32. [32]

    Toward an Ising model of can- cer and beyond

    Salvatore Torquato. Toward an Ising model of can- cer and beyond. Physical Biology, 8(1):015017, 2011

    work page 2011

  33. [33]

    Cellular Potts modeling of tumor growth, tumor invasion, and tumor evolution

    Andr´ as Szab´ o and Roeland MH Merks. Cellular Potts modeling of tumor growth, tumor invasion, and tumor evolution. Frontiers in oncology , 3:87, 2013

    work page 2013

  34. [34]

    The cellular Ising model: a framework for phase transitions in multi- cellular environments

    Marc Weber and Javier Buceta. The cellular Ising model: a framework for phase transitions in multi- cellular environments. Journal of The Royal Society Interface, 13(119):20151092, 2016

    work page 2016

  35. [35]

    Statistical Physics, volume 5 of Course of Theoretical Physics

    Lev D Landau and Evgenii M Lifshitz. Statistical Physics, volume 5 of Course of Theoretical Physics. Pergamon Press, 1958

    work page 1958

  36. [36]

    On Ising’s model of ferromag- netism

    Rudolf E Peierls. On Ising’s model of ferromag- netism. Mathematical Proceedings of the Cambridge Philosophical Society, 32(3):477–481, 1936

    work page 1936

  37. [37]

    Time Series Analysis: Forecasting and Control

    George E P Box and Gwilym M Jenkins. Time Series Analysis: Forecasting and Control . Holden- Day, 1970

    work page 1970

  38. [38]

    Magnetisation and mean field theory in the Ising model

    Dalton A R Sakthivadivel. Magnetisation and mean field theory in the Ising model. SciPost Physics Lec- ture Notes, 35, 2022

    work page 2022

  39. [39]

    Hopfield Networks is All You Need

    Hubert Ramsauer, Bernhard Sch¨ afl, Johannes Lehner, Philipp Seidl, Michael Widrich, Thomas Adler, Lukas Gruber, Markus Holzleitner, Mi- lena Pavlovi´ c, Geir Kjetil Sandve, et al. Hop- field networks is all you need. arXiv preprint arXiv:2008.02217, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2008

  40. [40]

    Univer- sal hopfield networks: A general framework for single-shot associative memory models

    Beren Millidge, Tommaso Salvatori, Yuhang Song, Thomas Lukasiewicz, and Rafal Bogacz. Univer- sal hopfield networks: A general framework for single-shot associative memory models. In Inter- national Conference on Machine Learning , pages 15561–15583. PMLR, 2022

    work page 2022

  41. [41]

    A new frontier for Hopfield net- works

    Dmitry Krotov. A new frontier for Hopfield net- works. Nature Reviews Physics, 5(7):366–367, 2023

    work page 2023

  42. [42]

    Morphogenesis and Evol- ution

    Keith Stewart Thomson. Morphogenesis and Evol- ution. Oxford University Press, 1988

    work page 1988

  43. [43]

    Multiscale modeling of form and function

    Adam J Engler, Patrick O Humbert, Bernhard Wehrle-Haller, and Valerie M Weaver. Multiscale modeling of form and function. Science, 324(5924):208–212, 2009

    work page 2009

  44. [44]

    Morphogenetic fields in embryogen- esis, regeneration, and cancer: non-local control of complex patterning

    Michael Levin. Morphogenetic fields in embryogen- esis, regeneration, and cancer: non-local control of complex patterning. Biosystems, 109(3):243–261, 2012

    work page 2012

  45. [45]

    Coordination of morphogenesis and cell- fate specification in development

    Chii J Chan, Carl-Philipp Heisenberg, and Takashi Hiiragi. Coordination of morphogenesis and cell- fate specification in development. Current Biology, 27(18):R1024–R1035, 2017

    work page 2017

  46. [46]

    Morphogenesis as Bayesian infer- ence: A variational approach to pattern formation and control in complex biological systems

    Franz Kuchling, Karl Friston, Georgi Georgiev, and Michael Levin. Morphogenesis as Bayesian infer- ence: A variational approach to pattern formation and control in complex biological systems. Physics of Life Reviews , 33:88–108, 2020

    work page 2020

  47. [47]

    Collective in- telligence: A unifying concept for integrating bio- logy across scales and substrates

    Patrick McMillen and Michael Levin. Collective in- telligence: A unifying concept for integrating bio- logy across scales and substrates. Communications Biology, 7(1):378, 2024

    work page 2024

  48. [48]

    Integrated in- formation in discrete dynamical systems: motiva- tion and theoretical framework

    David Balduzzi and Giulio Tononi. Integrated in- formation in discrete dynamical systems: motiva- tion and theoretical framework. PLoS Computa- tional Biology, 4(6):e1000091, 2008

    work page 2008

  49. [49]

    Functional specificity in the hu- man brain: a window into the functional archi- tecture of the mind

    Nancy Kanwisher. Functional specificity in the hu- man brain: a window into the functional archi- tecture of the mind. Proceedings of the National Academy of Sciences, 107(25):11163–11170, 2010

    work page 2010

  50. [50]

    Interactive specialization: a domain-general framework for human functional brain development? Developmental Cognitive Neur- oscience, 1(1):7–21, 2011

    Mark H Johnson. Interactive specialization: a domain-general framework for human functional brain development? Developmental Cognitive Neur- oscience, 1(1):7–21, 2011

    work page 2011

  51. [51]

    Neural and phenotypic representation un- der the free-energy principle

    Maxwell Ramstead, Casper Hesp, Alexander Tschantz, Ryan Smith, Axel Constant, and Karl Friston. Neural and phenotypic representation un- der the free-energy principle. Neuroscience & Biobe- havioral Reviews, 120:109–122, 2021

    work page 2021

  52. [52]

    Spontan- eous motion in hierarchically assembled active mat- ter

    Tim Sanchez, Daniel T N Chen, Stephen J DeCamp, Michael Heymann, and Zvonimir Dogic. Spontan- eous motion in hierarchically assembled active mat- ter. Nature, 491(7424):431–434, 2012

    work page 2012

  53. [53]

    Do hier- archical structures assemble best via hierarchical pathways? Soft Matter, 9(29):6851–6861, 2013

    Thomas K Haxton and Stephen Whitelam. Do hier- archical structures assemble best via hierarchical pathways? Soft Matter, 9(29):6851–6861, 2013

    work page 2013

  54. [54]

    Hierarchical assembly may be a way to make large information-rich structures

    Stephen Whitelam. Hierarchical assembly may be a way to make large information-rich structures. Soft Matter, 11(42):8225–8235, 2015

    work page 2015

  55. [55]

    Dispersity effects in polymer self-assemblies: a mat- ter of hierarchical control

    Kay E B Doncom, Lewis D Blackman, Daniel B Wright, Matthew I Gibson, and Rachel K O’Reilly. Dispersity effects in polymer self-assemblies: a mat- ter of hierarchical control. Chemical Society Re- views, 46(14):4119–4134, 2017

    work page 2017

  56. [56]

    Active materials: minimal mod- els of cognition? Adaptive Behavior, 28(6):441–451, 2020

    Patrick McGivern. Active materials: minimal mod- els of cognition? Adaptive Behavior, 28(6):441–451, 2020

    work page 2020

  57. [57]

    Competency in navigating arbitrary spaces as an invariant for ana- lyzing cognition in diverse embodiments

    Chris Fields and Michael Levin. Competency in navigating arbitrary spaces as an invariant for ana- lyzing cognition in diverse embodiments. Entropy, 24(6):819, 2022

    work page 2022

  58. [58]

    Darwin’s agential materials: evol- utionary implications of multiscale competency in developmental biology

    Michael Levin. Darwin’s agential materials: evol- utionary implications of multiscale competency in developmental biology. Cellular and Molecular Life Sciences, 80(6):142, 2023

    work page 2023

  59. [59]

    Machine wanting

    Daniel W McShea. Machine wanting. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomed- ical Sciences, 44(4):679–687, 2013

    work page 2013

  60. [60]

    Living things are not (20th century) machines: updating mech- anism metaphors in light of the modern science of machine behavior

    Joshua Bongard and Michael Levin. Living things are not (20th century) machines: updating mech- anism metaphors in light of the modern science of machine behavior. Frontiers in Ecology and Evolu- tion, 9:650726, 2021

    work page 2021

  61. [61]

    Rage against the what? the machine metaphor 11 in biology

    Ann-Sophie Barwich and Matthew James Rodrig- uez. Rage against the what? the machine metaphor 11 in biology. Biology & Philosophy , 39(4):14, 2024

    work page 2024

  62. [62]

    Formal thought disorders: from phenomenology to neurobiology

    Tilo Kircher, Henrike Br¨ ohl, Felicitas Meier, and Jennifer Engelen. Formal thought disorders: from phenomenology to neurobiology. The Lancet Psy- chiatry, 5(6):515–526, 2018

    work page 2018

  63. [63]

    A brief history of stigmergy

    Guy Theraulaz and Eric Bonabeau. A brief history of stigmergy. Artificial life, 5(2):97–116, 1999

    work page 1999

  64. [64]

    Stigmergic epi- stemology, stigmergic cognition

    Leslie Marsh and Christian Onof. Stigmergic epi- stemology, stigmergic cognition. Cognitive Systems Research, 9(1-2):136–149, 2008

    work page 2008

  65. [65]

    Bacterial stigmergy: an organising principle of multicellular collective behaviours of bacteria

    Erin S Gloag, Lynne Turnbull, and Cynthia B Whitchurch. Bacterial stigmergy: an organising principle of multicellular collective behaviours of bacteria. Scientifica, 2015(1):387342, 2015

    work page 2015

  66. [66]

    Stigmergy as a universal coordin- ation mechanism I: Definition and components.Cog- nitive Systems Research, 38:4–13, 2016

    Francis Heylighen. Stigmergy as a universal coordin- ation mechanism I: Definition and components.Cog- nitive Systems Research, 38:4–13, 2016

    work page 2016

  67. [67]

    Statistical mechanics of neural networks near saturation

    Daniel J Amit, Hanoch Gutfreund, and Haim Som- polinsky. Statistical mechanics of neural networks near saturation. Annals of Physics , 173(1):30–67, 1987

    work page 1987

  68. [68]

    Mean-field message-passing equa- tions in the Hopfield model and its generalizations

    Marc M´ ezard. Mean-field message-passing equa- tions in the Hopfield model and its generalizations. Physical Review E, 95(2):022117, 2017

    work page 2017

  69. [69]

    Theory of spin glasses

    Samuel Frederick Edwards and Phil W Anderson. Theory of spin glasses. Journal of Physics F: Metal Physics, 5(5):965, 1975. 12

    work page 1975